Histogram examples

The purpose of this utility package is to give full access to the customization and calibration of color ramps in the R environment. The resulting objects can be instantly passed to various plotting methods, such those provided by sf, terra and fields to mention a few. See some examples below and on the package’s website: https://adamtkocsis.com/rampage/.

library(rampage)

The package itself relies on grDevices::colorRampPalette function, which is used to linearly interpolate between colors that are tied to specific values.

The basic functionality of the package can be combined with any colors, either custom-selected or coming from a color ramp function. To provide an example, rampage includes a simple 5-color ramp called gradinv:

gradinv
#> function (n) 
#> {
#>     x <- ramp(seq.int(0, 1, length.out = n))
#>     if (ncol(x) == 4L) 
#>         rgb(x[, 1L], x[, 2L], x[, 3L], x[, 4L], maxColorValue = 255)
#>     else rgb(x[, 1L], x[, 2L], x[, 3L], maxColorValue = 255)
#> }
#> <bytecode: 0x5615cb54cbe0>
#> <environment: 0x5615cb5a08a0>

Histograms

For the sake of a minimal demonstration, rampage can be used to colorize bars of a histogram depending on their domain (breakpoints). We can illustrate this with sample from a Gaussian distribution:

set.seed(1)
vals<- rnorm(50000,1,1)

Color tiepoint data frame

The definition of color ramps depend on the given colors ("color") and the the values they are tied to. In the simplest case, we can use

df <- data.frame(
  z=c(-4, 1,  6),
  color=gradinv(3)
)
df
#>    z   color
#> 1 -4 #33358A
#> 2  1 #FFF99A
#> 3  6 #690720

Using the default z and color arguments:

ex <- expand(df, n=50)

This hist function can be used with a set of breakpoints (breaks) that are used to define the bars. The color of the individual bars is vectorized, if exactly as many colors as many bars are provided (number of breaks minus one), then the bars will be colored accordingly.

hist(vals, breaks=ex$breaks, col=ex$col, border=NA)

More breaks

The number of levels (breakpoints -1) is directly controled by n. A finer resolution can be achieved with

exMore <- expand(df, n=100)
hist(vals, breaks=exMore$breaks, col=exMore$col, border=NA)

More tiepoints

Specifying more tiepoints in the data.frame will give more control over the exact placement of the colors. For example if we want to clearly separate the values around 1 (population mean), then we can tie more colors to values that just above and below the values

dfMoreTies <- data.frame(
  z=c(-4, 0.5, 1,  1.5, 6),
  color=gradinv(5)
)
dfMoreTies
#>      z   color
#> 1 -4.0 #33358A
#> 2  0.5 #76ACCE
#> 3  1.0 #FFF99A
#> 4  1.5 #E22C28
#> 5  6.0 #690720

This can then be expanded to a full color ramp:

exMoreTies <- rampage::expand(dfMoreTies, n=200)

Which can be visualized with the hist function.

hist(vals, breaks=exMoreTies$breaks, col=exMoreTies$col, border=NA)

The exact placement of the colors are easily customizable by tweaking the z values in the data.frame that is expanded into the ramp. For instance, if we want the colorings to reflect the values (e.g. 0 is more important than the mean):

dfSkewed <- data.frame(
  z=c(-4, -1, 0,  1, 6),
  color=gradinv(5)
)
dfSkewed
#>    z   color
#> 1 -4 #33358A
#> 2 -1 #76ACCE
#> 3  0 #FFF99A
#> 4  1 #E22C28
#> 5  6 #690720

Expanding to a full color ramp:

exSkewed <- rampage::expand(dfSkewed, n=200)
hist(vals, breaks=exSkewed$breaks, col=exSkewed$col, border=NA)

See the package website for package-specific examples.